divakarwl12

@divakarwl12 source repo

6 skills
0 followers
1 day ago last updated

▌
Promo Video · divakarwl12 bundle
Create professional promotional videos using Remotion with AI voiceover and background music. Invoke with /promo-video.

0
installs
▌
Creating Shot List · divakarwl12
Reads story.md, characters.json, and backgrounds.json to decompose the story into individual 8-second shots. Each shot maps to specific character IDs, a background ID, camera movement, action description, and a self-contained Veo 3 video-generation prompt. Saves output as shots.json. Part of the Story-to-Animation pipeline (Step 4 of 5). Use when story and asset JSON files exist and the user wants to create the shot-by-shot breakdown for video generation. After generating shots.json, ALWAYS present the output and wait for explicit user approval. Do NOT automatically trigger the next pipeline step.

0
installs
▌
Generating Story From Logline · divakarwl12
Expands a short story logline or premise into a full structured narrative screenplay saved as story.md. Use when the user provides a story idea, logline, or premise and wants to generate a complete story with scenes, characters, dialogue, and visual descriptions for 3D animation production. Part of the Story-to-Animation pipeline (Step 1 of 5). After generating story.md, ALWAYS present the output and wait for explicit user approval. Do NOT automatically trigger the next pipeline step.

0
installs
▌
Generating Composite And Video · divakarwl12 bundle
Reads shots.json and processes each shot in two fully-parallel phases: (1) generates composite scene images by blending character and background images (imgbb image_url fields) via the kie.ai flux-2/pro-image-to-image API, saving to ./composites/; (2) generates video clips from each composite using the kie.ai sora-2-image-to-video API, saving to ./clips/. Both phases use the taskId-based /jobs/recordInfo polling so all shots run concurrently — total time is the slowest single job, not the sum. After clips are approved, runs scripts/merge_clips.py (FFmpeg) to concatenate all clips in narrative order into ./final_animation.mp4. API keys loaded from config.json. Part of the Story-to-Animation pipeline (Step 5 of 5, final step). Use when shots.json exists and characters.json/backgrounds.json have image_url fields (populated by generate_images.py in Step 3). After generating and merging, ALWAYS present output and wait for explicit user approval.

0
installs
▌
Extracting Characters And Backgrounds · divakarwl12
Reads story.md and extracts all unique characters and background locations. Generates detailed Pixar-style 3D image-generation prompts for each, saving them as characters.json and backgrounds.json. Part of the Story-to-Animation pipeline (Step 2 of 5). Use when story.md exists and the user wants to prepare character and background prompts for AI image generation. After generating both JSON files, ALWAYS present the output and wait for explicit user approval. Do NOT automatically trigger the next pipeline step.

0
installs
▌
Generating Character And Background Images · divakarwl12 bundle
Reads characters.json and backgrounds.json, then generates reference images for each character and background using the kie.ai nano-banana-pro API. Character images (1:1 aspect ratio) are saved to ./characters/ and background images (16:9) to ./backgrounds/, named by their IDs (char_001.png, bg_001.png, etc.). Each image is then uploaded to imgbb for permanent public hosting, and the imgbb URL is written back as an image_url field in both JSON files — these URLs are required by generate_videos.py. Part of the Story-to-Animation pipeline (Step 3 of 5). Run scripts/generate_images.py to execute. Use when both JSON prompt files exist and the user wants to generate the actual reference images. After generating all images, ALWAYS present the output and wait for explicit user approval. Do NOT automatically trigger the next pipeline step.

0
installs

6 published skills